Fpdf2

Fixed
- a "fake duplicates" bug when a `Pillow.Image.Image` was passed to `FPDF.image`

Added
- new features: **document outline & table of contents**! Check out the new dedicated [documentation page](https://py-pdf.github.io/fpdf2/DocumentOutlineAndTableOfContents.html) for more information
- new method `FPDF.text_annotation` to insert... Text Annotations
- `FPDF.image` now also accepts an `io.BytesIO` as input
Fixed
- `write_html`: properly handling `<img>` inside `<td>` & allowing to center them horizontally

Added
- `FPDF.set_xmp_metadata`
- made `<li>` bullets & indentation configurable through class attributes, instance attributes or optional method arguments, _cf._ [`test_customize_ul`](https://github.com/py-pdf/fpdf2/blob/2.3.2/test/html/test_html.py#L242)
Fixed
- `FPDF.multi_cell`: line wrapping with justified content and unicode fonts, _cf._ [118](https://github.com/py-pdf/fpdf2/issues/118)
- `FPDF.multi_cell`: when `ln=3`, automatic page breaks now behave correctly at the bottom of pages

Added
- `FPDF.polyline` & `FPDF.polygon` : new methods added by uovodikiwi - thanks!
- `FPDF.set_margin` : new method to set the document right, left, top & bottom margins to the same value at once
- `FPDF.image` now accepts new optional `title` & `alt_text` parameters defining the image title
and alternative text describing it, for accessibility purposes
- `FPDF.link` now honor its `alt_text` optional parameter and this alternative text describing links
is now properly included in the resulting PDF document
- the document language can be set using `FPDF.set_lang`
Fixed
- `FPDF.unbreakable` so that no extra page jump is performed when `FPDF.multi_cell` is called inside this context
Deprecated
- `fpdf.FPDF_CACHE_MODE` & `fpdf.FPDF_CACHE_DIR` in favor of a configurable new `font_cache_dir` optional argument of the `fpdf.FPDF` constructor

Many thanks to [eumiro](https://github.com/py-pdf/fpdf2/pulls?q=is%3Apr+author%3Aeumiro) & [fbernhart](https://github.com/py-pdf/fpdf2/pulls?q=is%3Apr+author%3Aeumiro) for their contributions to make `fpdf2` code cleaner!
Added
- `FPDF.unbreakable` : a new method providing a context-manager in which automatic page breaks are disabled.
_cf._ https://py-pdf.github.io/fpdf2/PageBreaks.html
- `FPDF.epw` & `FPDF.eph` : new `property` methods to retrieve the **effective page width / height**, that is the page width / height minus its horizontal / vertical margins.
- `FPDF.image` now accepts also a `Pillow.Image.Image` as input
- `FPDF.multi_cell` parameters evolve in order to generate tables with multiline text in cells:
* its `ln` parameter now accepts a value of `3` that sets the new position to the right without altering vertical offset
* a new optional `max_line_height` parameter sets a maximum height of each sub-cell generated
- new documentation pages : how to add content to existing PDFs, HTML, links, tables, text styling & page breaks
- all PDF samples are now validated using 3 different PDF checkers
Fixed
- `FPDF.alias_nb_pages`: fixed this feature that was broken since v2.0.6
- `FPDF.set_font`: fixed a bug where calling it several times, with & without the same parameters,
prevented strings passed first to the text-rendering methods to be displayed.
Deprecated
- the `dest` parameter of `FPDF.output` method

Added
- new unit tests, a code formatter (`black`) and a linter (`pylint`) to improve code quality
- new boolean parameter `table_line_separators` for `write_html` & underlying `HTML2FPDF` constructor
Changed
- the documentation URL is now simply https://py-pdf.github.io/fpdf2/
Removed
- dropped support for external font definitions in `.font` Python files, that relied on a call to `exec`
Deprecated
- the `type` parameter of `FPDF.image` method
- the `infile` parameter of `Template` constructor
- the `dest` parameter of `Template.render` method